In the 60s and 70s, almost any concert devoted to the music of Havergal Brian - not that there were many of those - could, almost by definition, be expected to include a world premiere. After all, he wrote so much music, and so little of it had been performed before his death in 1972. Nowadays the situation is gratifyingly different. Brian’s output is becoming accepted as a major and unavoidable phenomenon of 20th century British musical life - witness his coverage in BBC Radio 3’s recently published book of British music Fairest isle, and the fact that nearly a dozen of Brian’s symphonies were broadcast in the morning anthology programme Musical encounters. In fact, all 32 of Brian’s symphonies have been performed - some of them in public, a lot of them more than once; more than half of them have been commercially recorded and the current Marco Polo cycle will include all 32.

However, the situation with his five operas, which he valued as highly as, and in some cases higher than, his symphonies, is rather different. None of them has ever been staged. His first, The Tigers, and his last, Agamemnon, have both been broadcast complete in studio recordings by the BBC. But of the other three, only the prologue of the opera based on the first part of Goethe’s Faust has been performed, and of the opera based on Shelley’s The Cenci, only the overture and a couple of fanfares have been heard [MM is writing in 1995; in 1997 the HBS mounted the premiere, in concert, of The Cenci].

A nd as for the remaining opera, Turandot, Prinzessin von China (Turandot, Princess of China), until the HBS 21st Anniversary Concert the only scrap of it brought before the public was a pair of piano arrangements I made over 20 years ago of the Minuet and Lugubre-Marsch, which have been occasionally performed and broadcast. Not a bar of Brian’s quite massive orchestral score had ever been played until this year’s concert, which featured the world premiere of no less than nine orchestral extracts from this very rich operatic work.

So, I’d like to say something about Brian’s Turandot, and the justification for taking extracts from it. But one can’t really discuss Brian’s Turandot without considering it in relation to the other Turandots, especially perhaps the Turandot, the opera on the subject with which audiences are most familiar. No, Brian’s Turandot, though it contains much taxing and characteristic vocal writing, can boast no great Romantic arias in the manner of In questa reggia or Nessun dorma!. It’s a very different kind of work: as he subtitled it, Ein tragikömisches Märchen - a tragi-comic fairytale.

The 20th century has seen, in fact, three great operas on the Turandot story. The first is Busoni’s two-acter, in German, composed in 1916 and 1917 and founded upon the orchestral suite he’d already written in 1905 and incidental music in 1911. Then comes Puccini’s, in three acts and in Italian, of course - his last opera, toiled over from 1921 to his death in 1924 and left unfinished. And finally, Havergal Brian’s Turandot, also in three acts like the Puccini, but in German like the Busoni, which he began composing late in 1949 and completed on 18 May 1951. He was 75, nearly 10 years older than Puccini had been when he died, and Turandot was only the second opera he had completed, though at least two others had been begun and several more contemplated at various times. Yet in the next 17 years, up to the age of 92, Brian was still to compose three further operas and 24 symphonies, as well as other works.

Now these three very different operas - Busoni, Puccini, Brian - are all ultimately based on Turandot (or Turandotte), a five-act fairytale play, a fiabe chinese, as he called it, written in 1761 or 1762 depending on which source you follow, by the 18th-century Venetian dramatist and diplomat, Count Carlo Gozzi. And thereby hangs a very complicated tale.

Gozzi, though an amateur playwright, is a key figure in the development of the Italian theatre. He’s perhaps best remembered for his attempted revival of the Commedia dell’Arte - the comic tradition, going back to the travelling acting troupes of 16th century Italy, which largely improvised farces set around stock situations and stock characters wearing masks - characters such as Pantaleone, the grumpy old guardian, and Columbina, the clever maid beloved by the romantic hero, and so on.

I mention these two particularly, of course, because of one of HB’s very earliest works, Pantalon and Columbine, which reminds us that these stock characters had by the 19th century become part of the common literary heritage of Europe, and of this country, permeating English pantomime and even seaside entertainment: Mr. Punch is another Commedia dell’Arte figure, Pulcinella. Harlequin - Arlecchino - is another obvious example.

In time, the Commedia dell’Arte became a kind of interlude in more literary comedies: and by the mid-18th century, on the Venetian stage, it was in retreat before the reforms of Goldoni, who was substituting his own more realistic characters and dialogue - master/servant intrigue, social comment - for the stock figures and their improvised comedy. When Gozzi, who wanted to preserve the realm of fantasy and imagination with the Commedia dell’Arte, attacked Goldoni, the latter claimed that his enormous public success proved the desirability of his innovations.

Gozzi’s rejoinder was that the public was so fickle, it would applaud even if a children’s fairytale was put on the stage. And to prove his point, he went on to write his ten fiabe, fairy-tales, masterpieces of fantasy and grotesque - in which the Commedia dell’Arte figures once more assume their former function of improvisatory interludes in the action, but also stand in a new, deliberately anachronistic relationship to it: they are in the action yet not of it, talking in Venetian dialect while the other characters speak literary Italian, and making jokes and topical allusions about contemporary Venetian society. They are Venetian low-life inserted into romantic or legendary or mythical settings, mediating and subverting.

One opera - not actually based on Gozzi, but illustrating the principle very nicely - is Richard Strauss’s Ariadne auf Naxos, with its astonishing mixture of a quartet of Commedia dell’Arte characters (Zerbinetta is a kind of Columbina) into the world of Classical myth; Classical myth as the 18th century understood it, as Molière’s Bourgeois Gentilhomme understood it, that is.

This combination of the fantastic and the bizarre in Gozzi’s plays made them very attractive to the 19th century Romantics, and two of his other fiabe have inspired notable 20th century opera L'amore delle tre melarance became Prokofiev’s Love of three oranges, while Il re Corvo gave Hans Werner Henze the subject of his König Hirsch (King Stag). Not quite in passing, it’s interesting to recall that Il re Corvo is one of the very few literary works to have the distinction of having been actively considered for an opera libretto by - Brahms.

Now Brahms, as we all know, never wrote an opera, not least perhaps because of the overwhelming predominance of Wagner on the operatic stage of his time. But he certainly thought about doing so, and the idea of a drama by Gozzi prompted him to speculate on the possibility of a kind of opera so different from Wagner that there would essentially be no grounds for comparison. A Brahmsian opera would be almost a different art-form; something essentially light, perhaps not even through-composed: a comedy with serious undertones, a kind of singspiel with musical heightening at important moments, more like The magic flute than The ring of the Nibelung. In fact The magic flute’s mixture of high and low characters, high and low idioms, in a fantastic setting, has a lot in common with the plays of Gozzi.

But to return to Turandot. The legend of Turandot - or at least, of the princess who must be married to an unknown prince - is actually of Persian origin. In the 13th century the Persian poet Lari transferred the action of the tale to China, and it seems to be he who first used the name by which the princess is known, from Turan, the Persian word for China, and doht, meaning daughter: thus ‘daughter of China’. In due course the story became one of the collection of A thousand and one nights, introduced into Europe in translation about the end of the 17th century.

It is however Gozzi’s play which has chiefly been responsible for its further dissemination, especially as a subject for music. It’s from Gozzi that we derive the basic plot, which Havergal Brian’s opera follows, on the whole, very faithfully. Kalaf, son of King Timur who was defeated by the Emperor of China, has come incognito to Peking. Here he meets his former servant Barak, also living incognito, who tells him of Turandot, the Emperor’s beautiful daughter. Hating men, she sets each of her suitors three riddles; if they cannot answer them they are beheaded. The latest victim is the Prince of Samarkand, whose grief-stricken servant passes by and throws a portrait of Turandot to the ground. On seeing the portrait, Kalaf is enraptured and vows to undergo the trial of the riddles.

A t the trial, presided over by the old Emperor Altoum, Kalaf refuses to reveal his name, but correctly answers all the riddles. By her own rules Turandot must many him, but she refuses and threatens to kill herself. Kalaf offers her a riddle in return: if she can guess his name, his life will be forfeit, if not she must become his wife. But Kalaf has been recognised by Turandot’s servant Adelma, who knew him when - again incognito - he served her father King Keikobad. She tries to persuade Kalaf to escape with her, and tricks him into telling her his name. At the renewed trial, Turandot and her retinue arrive in mourning, but she then both proclaims his name and spares him. Kalaf tries to stab himself, but the Princess prevents him, and they declare their mutual love.

In what I said earlier I’ve perhaps over-stressed the importance of the Commedia dell’Arte characters, who as you’ll have noticed haven’t rated a mention in that very abbreviated plot-synopsis. But four Commedia dell’Arte figures do appear as courtiers: Tartaglia as the Emperor’s chamberlain, Pantalone as his secretary, Brighella as captain of the guard, and Truffaldino as chief eunuch; and they play an important if incidental part in Gozzi’s play. As Busoni wrote, ‘these masked figures, familiar to Italians, contribute to the effect of fantasy, for they throw a bridge from the Venetian public into the fictitious Orient of the stage, and in this way destroy the illusion that what is going on is real life’.

It is in fact Busoni’s Turandot which still allots the Commedia dell’Arte characters an important role (even though he reduces their number by cutting out Brighella). Puccini uses them least, or rather not at all. Havergal Brian actually uses all four of them (though the part of Truffaldino is merely mimed), but their function is less comically bizarre and subversive. So all three of the great 20th century Turandots are considerably altered from their prime source in Gozzi. The closest in general effect is Busoni: yet, though he wrote his own German libretto, based directly on Gozzi’s play, he drastically compressed the action, introducing certain new characters and plot-twists, and cut out (as all the Turandot operas do) certain obscene elements which seem to have been quite acceptable to 18th century Venice but not to later sensibilities.

But his light, ironic, fantastical tone has a good claim to be closest to Gozzi’s intentions. Ironically the opera which departs most radically from Gozzi’s original is the Puccini, the Italian one. The character of Liù, for instance, a rival heroine to Turandot herself, is an invention of Puccini’s librettists Simoni and Adami, though she has some of the features of Gozzi’s Adelma; the Commedia dell’Arte characters disappear, and their place is taken by those picturesque and sentimental stage-Chinamen Ping, Pang, and Pong. Essentially, though, this is the great Romantic Turandot, the work of colour and passion.

Brian’s opera stands somewhat between, and yet at an angle, to these different removes from Gozzi’s play. Like Busoni, he prepared his own libretto quite faithfully from his original; but his original was not the Gozzi itself, but an intermediate text - namely the free adaptation and translation which Friedrich Schiller made in 1801 for the Hoftheater in Weimar. (And here I should perhaps note the curious fact that Puccini was originally inspired not by the Gozzi Turandot but by the Schiller Turandot, for the first form in which he encountered the play was actually an Italian translation - or rather re-translation - of Schiller’s German. Confusing, isn’t it?).

In adapting Gozzi’s play for the German stage of a generation or so later, Schiller had, precisely, great problems with the anachronistic figures of the Commedia dell’Arte. His solution was to cut out the more indecent exchanges and bring Tartaglia et al somewhat up the social scale, to suppress their distinctive comic dialect, put them linguistically on the same level as the more elevated characters. The result is certainly that the humour becomes less functional and more sententious - more ponderous, even. Busoni, again, criticised Schiller’s version for ‘failing to convey the inherent feeling, even in scenes that border upon tragedy, that one is always involved with a matter of fantasy’.

Nevertheless, Schiller’s Turandot was itself an influential text of German Romanticism, and throughout the 19th century it was much more widely known than Gozzi’s. It was also connected with one of the early Romantic composers, namely Weber, for it was in 1809 that Weber wrote some rather good incidental music for a Weimar revival of Schiller’s play. And it was from Weber’s Turandot music that Paul Hindemith, in the 1940s, derived the subject of the ‘Turandot Scherzo’ in his well-known orchestral work Symphonic metamorphoses on themes of Weber.

So: why, in 1949, did Havergal Brian turn to the story of Turandot - and specifically Schiller’s Turandot - for the subject of his second opera? I’m not sure that be ever explained his reasons. He did say, I think, that he’d never heard the Puccini Turandot (which is quite possible) and that at the time be wrote his own opera he didn’t even know of the existence of the Busoni. Given that as recently as 1939 he’d written about Busoni’s Arlecchino, composed to go in a double bill with Turandot, it’s more likely that he’d forgotten it. He was certainly aware, in the inter-war period, of Busoni’s earlier incidental music to Gozzi’s play. But I don’t doubt that, as far as the text goes, it was the Schiller Turandot with which Brian was familiar, rather than the Gozzi.

As to why he should choose it in the first place, essentially we’re reduced to speculation - there are a number of pointers, but none of them conclusive.

1: A couple of years before, Brian had started writing an opera on JM Synge’s tragedy, Deirdre of the sorrows - but he’d had to abandon it when be found the publishers wouldn’t allow him copyright permission. That experience seems to have decided him that any future opera he attempted should be based on classic drama by authors historical enough to be safely out of copyright. This is the case with all four remaining operas that he composed between 1949 and 1957. Turandot is the first of these.

2: he may well have felt that a setting of a classic German-language drama would be open to much greater opportunities for production on the Continent than existed in the UK. Despite the examples of Rutland Boughton’s The immortal hour and Britten’s Peter Grimes, the idea of British opera was by no means firmly established, as Brian had first-hand experience from the vicissitudes of his own first opera, The Tigers. He was a passionate admirer of German Romantic literature, especially Goethe and Schiller: in 1955-56 he would eventually write an opera on Goethe’s Faust, also in German, and that drama had already inspired his Gothic Symphony in the 1920s. He may even have been attracted by the rather monumental side of Schiller’s Turandot, given his own taste for the monumental.

3: Brian’s music often, in various ways, shadows that of his hero, Richard Strauss, even if it hardly ever sounds like him. His early comedy overture, Doctor Merryheart, parodies the programmatic techniques of the Straussian symphonic poem. The Tigers contains direct parodies of Strauss. The Gothic Symphony is dedicated to Strauss whom he addresses as his ‘beloved friend and master’; and the last opera, Agamemnon, though composed in English, was designed as a curtain-raiser to Strauss’s Elektra.

Almost certainly, Brian was aware of the fact that Schiller’s Turandot had been a strong influence on Strauss’s Die Frau ohne Schatten, an opera he enormously admired. Perhaps he wanted to write a kind of Frau ohne Schatten - without the top-heavy structure of Hofmannsthal’s psychological symbolism, but with, let’s say, a similar kind of monumental exoticism, a similar story of mysterious trials overcome in a fantastic setting. A trial opera, like The Magic Flute, in whose tradition Die Frau ohne Schatten stands. Rather like the modern Magic flute which Brahms thought he might be able to fashion from one of Gozzi’s dramas. This I think is perhaps what most distinguishes Brian’s opera from those of Busoni and Puccini.

Because, 4: I wonder if it wasn’t precisely Hindemith’s Symphonic metamorphoses on themes of Carl Maria von Weber, first performed in this country at the Proms in 1946, which directed Brian, through its movement based on Weber’s Turandot incidental music, to consider Schiller’s Turandot as an opera text. Brian admired Hindemith, and would have taken an interest in any recent work of his. And in Brian’s Turandot there is a curious atmosphere that surfaces again and again, which I can only think of as a kind of ‘mock classicism’ - not neo-classicism but a kind of burlesque of Weimar classicism, of the classical style in music - of Weber, Beethoven even, which in its own way subtly undercuts the comparative solemnity of Schiller’s text.

This quality is very clear in, for instance, the Minuet in act two, the prelude to that act, and the Lugubre-Marsch, which is a kind of parody of the Funeral March in the Eroica Symphony, perhaps combined with that in the Chopin B flat minor sonata. Now you find hints of this ‘mock-classicism’ elsewhere in Brian, notably in the giganticised minuet measure that invades the finale of his ninth symphony, and another minuet-like episode in the central movement of his 11th. But it seems to be developed here into something quite specific to Turandot, making it a rather unique work in his output. Though Brian is making use of the full panoply of the 20th century orchestra and the full resources of post-Romantic harmony, there’s a formality about the piece that relates to earlier times - to far-off times.

A nd maybe this is the way Brian compensates for Schiller’s underplaying the fantastic element in his version of the drama. Because certainly, as I far as I can see and hear, he doesn’t imitate or quote oriental melodies as Puccini, Busoni, and indeed Weber do. Rather, his legendary China is perhaps Weimar of the 1800s, a far-off world of ancient times; after all, doubly far-off in the aftermath of the Second World War. The formality, the playing at being a classical composer, points up the work’s distant family connection with The magic flute, and also imparts a rich seam of ambiguity to the proceedings, anachronistic as the presence of the Masks: a pleasurable, ironic, distancing effect. I don’t mean to suggest that the opera is composed in a classical style: its basic style is that of Brian’s symphonic works of this period. But it’s a particular facet of that style which seems to me to bulk quite large.

This, anyway, is what I got from a study of the score. Only a complete performance of Havergal Brian’s Turandot will in fact reveal if I’m right. But you can feel it, I think, in the compendious orchestral extracts played at the May concert.

What justification is there, I can perhaps hear some ask, for presenting orchestral extracts? Well, probably all that people who know Brian’s output reasonably well need to know is that Turandot comes chronologically between his eighth and ninth symphonies, the eighth of 1949 and the ninth of 1951, two of his most inventive orchestral scores from one of his most productive periods - that extraordinary nine years, 1948-1957, which saw the production of, among other things, seven symphonies and four operas. You can certainly hear affinities between Turandot and Symphonies 8 and 9 - in fact I rather fancy you can hear, as the opera unfolds, a shift from the texturally multifarious and iridescent, rather filigree world of parts of No 8 towards the more massive effects characteristic of No 9.

But in any case the ultimate authority for presenting orchestral extracts from Turandot comes from Havergal Brian himself - for he it was who prepared the first three extracts for concert performance. Though the voices matter a great deal in Brian’s Turandot, of the three great 20th century operas on the subject it’s perhaps the one where they matter the least. The drama is essentially carried in the orchestra.

To get back to Gozzi for a moment; Gozzi was acutely aware of the role music had to play in creating the fantastic ambience of his plays, and in Turandot he specifies music of various kinds at different points in the drama. This was how Busoni’s Turandot came into being: he composed a concert suite based on Gozzi’s indications for music. Then he enlarged it into incidental music for an actual production of Gozzi’s play in Berlin. And finally he added voices and developed all the music he had so far composed into an opera. The play, then, is full of spaces for musical set-pieces. Not even Puccini ignores that feature - I’m thinking of things like the march which leads to the entrance of the Emperor in Puccini’s act 2 - a precise parallel in function and dramatic position to the march in the Prelude to Act Two of Brian’s opera.

Like his first opera The Tigers, Brian’s Turandot contains a large amount of music for orchestra alone, or where the contribution of the voices is fairly minimal. In The Tigers, that yielded a number of ‘symphonic dances’, as Brian called them, apt for concert performance. In Turandot the range of possible extracts is wider, from scene-painting and scene-changing, through dances and pageantry, to processionals. Brian himself was thoroughly aware of such possibilities. On 26 November 1962 - that’s over a decade since he had completed the opera - he wrote to Robert Simpson ‘I am making orchestral extracts from my Turandot: fine stuff this’. The task took him over four months, a rather longer period than he tended to need at that time for creating a new symphony. He’d apparently considered the idea of a four- or five-movement suite including voices, but what he eventually produced was a purely orchestral score which he just called Three pieces from Turandot. They all come from act 1 of the opera: in fact the first piece is the prelude to that act, and the other two run virtually in sequence after it.

Now it may well he that act 1 of Brian’s opera will eventually prove to be problematical on stage because, essentially, it consists of a single long conversation between Kalaf and Barak: mutual recognition, story so far, each filling in the other on what’s happened since they last met, Barak’s account of Turandot and the challenge to her suitors, Kalaf’s decision to dare to sue for her hand. Basically it’s a tenor and a baritone singing at each other over a long period - while, in the orchestra, a fantastic cornucopia of sonic invention and luminous detail is poured out, some of it depicting what is being narrated, some of it independent invention. Brian clearly felt this stuff was too good to risk being buried as a mere accompaniment for the whole of its existence. So he made this separate score - the Three pieces from Turandot - preserving that orchestral texture verbatim, except for a few phrases from the vocal parts transferred to the wind instruments.

Another decade passes, and I come into the picture. I’d seen the score Turandot at the BBC, and I’d seen the score of the Three pieces. I’d wondered why Brian selected his extracts only from act 1, when the other two acts seemed equally rich, if not richer, in music that cried out to be heard in the concert hall. I wrote to him about it - this was in the penultimate year of his life, and he had ceased composing entirely, but I suggested that perhaps another group of extracts would be the kind of light task he might be able to contemplate.

He replied to me on 13 September 1971, from Shoreham-by-Sea: ‘About the extracts from operas, I agree with what you say of them and I will think about your suggestion re Turandot... when I can get the full scores. ... At my advanced age [he was 95, of course] health varies very much & to do fine creative work one must be in first class health condition. Such variation in health & sight do not inspire creative work. Anyhow I will think about extracts from the operas’. In a later letter, sent on 2 October, be recalled that he had arranged the Three pieces with Bryan Fairfax and his Polyphonia Orchestra in mind, because Fairfax had seen the opera in score and had been very impressed by it. Fairfax and the Polyphonia had of course premiered Brian’s Symphony No 18 earlier in 1962, before he started arranging the Three pieces: but these call for a considerably larger orchestra than the symphony does, and whether Brian ever apprised Fairfax of the existence of the Turandot pieces I simply don’t know.

Of course, when I wrote to him, the full score Turandot was lodged with the BBC, and as far as I know he never saw it again before he died, so that was that. But since, in principle, he seemed to have conceded that further extracts might indeed be possible, I felt emboldened enough after his death, first to arrange two extracts for solo piano (the Minuet from act 2 and the Lugubre-Marsch from act 3), and later to devise a six-movement suite of music from acts 2 and 3, interfering no further with Brian’s scoring than he himself had done when he made his score of the Three pieces.

I do want to make it clear, however, that the fact that all nine movements were performed together then [say 1995] doesn’t mean that they can’t now be set asunder. Brian intended his Three pieces to be a self-sufficient concert item; I intended the same for the suite. Either can be performed, depending the programme time required to be filled as well as the availability of instruments (the suite does need a few more instruments than the Three pieces). But if all nine extracts are performed together, as they were on 27 May, then at least one gets a kind of orchestral synopsis of the course of the drama. You get the preludes to the three acts, passages of dialogue, a little dance, a nocturne, a scene-change, a big balletic entrée, and a march. And out of all this you may start to get a pretty good idea of the particular flavour and, indeed, magic of this fascinating opera.

NL120-1 / © Malcolm MacDonald 1995

BACK TO TITLE PAGE  

990326 Havergal Brian - the official website HOME